Correspondência verbal/não verbal: efeitos da intermitência de reforçamento e da história de fracasso escolar

Abstract

The present proposal is based on the assumption that verbal responding is operant behavior established and maintained by reinforcement contingencies, and that this behavior can be analysed as a dependent variable. Recent research about say-do correspondence have used the educational context to investigate conditions able to guarantee corresponding reports of errors and correct responses about academical activities. The present study investigated the influence of a history of school failure and task difficulty on self-report of two groups of children about their reading performance, as well the effects of reinforcement schedules on the establishment and maintenance of accurate reports. Four children with history of school failure and four children without such history, all of them attending Fundamental Education, participated. Doing consisted in reading aloud a written word presented on the computer s screen and Saying in reporting if the reading was correct or not. The experiment began with a General Pretest that evaluated the participants reading repertoire, followed by baseline sessions that verified frequency of corresponding reports as a function of increasing probability of errors in a session (10, 25, 50 e 70%). On the next phase, correspondence training sessions, in which corresponding reports were reinforced continuously and later intermittently (variable ratio), were conducted. Finally, a baseline return was conducted to evaluate if the previously trained conditions were sufficient to promote accurate self-report maintenance. Points exchangeable for gifts were used as consequences. There were no pronounced differences on the average of corresponding reports of errors or correct responses between the two groups during the first baseline, along the different error probabilities programmed (except during 50% error probability). On the other hand, most non-corresponding reports of error occurred when the programmed error probability was about 50% for the children with history of school failure and about 25% for those without such history. The correspondence training was effective to produce corresponding reports for all participants. The high correspondence level was maintained during the intermittent reinforcement phase as well during the subsequent extinction condition. Error frequency seemed to be a relevant variable for self-report accuracy for all children on this study, independently of the presence or absence of history of school failure and of the increasing levels of task difficulty. The intermittent reinforcement training was effective to maintain corresponding reports of errors and correct responses during subsequent conditions in which no consequences for responding were programmed.